Abstract: Various examples are directed to systems and methods for obscuring private information in input data. A system may apply an encoder model to an input data unit to generate a latent space representation of the input data unit. The system may apply multi-dimensional noise to the latent space representation of the input data unit, the multi-dimensional noise having a first value in a first latent space dimension and a second value different than the first value in a second latent space dimension. The system may apply a decoder model to the latent space representation of the input data unit to generate an obscured data unit.

Abstract: A method includes receiving by a chat server a client request from a client device communicating with the chat server. The chat server generates a server response that is transmitted to the client device. A nudge repository is searched for a nudge action based on a set of tokens generated from at least a portion of the client request. In response to finding the nudge action, a user cohort to receive the nudge action is determined. A nudge request including the nudge action and the user cohort is generated and transmitted to the chat server. The nudge action is deployed from the chat server to one or more client devices associated with one or more user identifiers in the user cohort. The nudge action is rendered as a nudge at the one or more client devices.

Abstract: A user experience repository may contain base layouts and variant metadata for applications of an enterprise. An application design platform may receive, from a designer, an indication of a selected base layout for a selected application and interact with the designer to create a user experience variant (e.g., a page layout). The designer may then define an assignment rule for the user experience variant, the assignment rule including custom logic and multiple user parameters (e.g., a user role, country, language, etc.), and the system may store information about the user experience variant and assignment rule. An enterprise application service platform may determine that a user is accessing the selected application and evaluate the custom logic of the assignment rule based on user parameters of the user accessing the selected application. In accordance with the evaluation, the system may arrange to provide the appropriate user experience variant to the user.

Abstract: Computer-readable media, methods, and systems are disclosed for processing backup and recovery of pages in an in-memory database system employing tenant-based, group-level encryption for a plurality of tenants. Page metadata for a page may be collected. The page may comprise a header and a set of page contents. The page metadata may be stored with the page header. When a backup request is received, the page may be loaded into a backup stream. The page may be loaded with the header unencrypted and the page contents encrypted. When a recovery request is received, the data page may be retrieved from the backup stream. A converter may be rebuilt using the page metadata. The data page may then be written to the main memory of the database system. The page contents may be decrypted by the converter.

Abstract: The present disclosure involves systems, software, and computer implemented methods for processing correlated calculated fields in correlated subqueries. One example method includes receiving a request to process a database query that includes at least one subquery and processing each subquery. A determination is made as to whether each subquery includes a correlated calculated field that is a calculated field that is not included in a first datasource queried by the subquery but is included in a second datasource queried by an outer query that includes the subquery. If the subquery does not include a correlated calculated field, the subquery is processed without modification. If the subquery includes a correlated calculated field, metadata is retrieved for the outer query of the subquery that queries the second datasource and includes the correlated calculated field. The subquery is rewritten using the metadata and the rewritten subquery is processed.

Abstract: Techniques and solutions are provided for improving query execution. Data models can be complex, which is often reflected in queries against such data models. The present disclosure provides a query refactoring technique where a complex query, such as a query expressed as a single select statement, can be formulated as a series of less complex queries. The workload of a database can be reduced by combining results of the less complex queries outside of the database. The present disclosure provides a framework for implementing these techniques, where the framework includes a virtual cube, a calculation engine, and one or more operations, which can all be implemented as classes in a programming language, and where a generic class or interface can help guide users in developing subclasses that provide a reformulation or refactoring of a complex query.

Abstract: Techniques and solutions are provided for improved data modelling, including techniques that facilitate data model comparison, integration, or deployment. A core data model is created that has a plurality of elements, which can represent semantic concepts, including relationships between semantic concepts. A plurality of domain models are mapped to core data model elements. In particular, when a domain model element is to be created, the element is mapped to a core data model element. When a relation is to be established between two domain model elements, the relation is mapped to a relation type represented in the core data model. Mapping to a core data model can be enforced as part of domain model creation, allowing different domain models to be compared, since domain elements in different domain models that represent a common semantic will be mapped to the same core data model element.

Abstract: Some embodiments provide a non-transitory machine-readable medium that stores a program executable by at least one processing unit of a device. The program retrieves a plurality of sets of master data from a plurality of master data sources. Each set of master data in the plurality of sets of master data includes a set of entities. For each entity in each set of entities, the program further determines whether the entity is consistent or inconsistent. The program also receives, from an application, a request for a set of entities. The program further determines a subset of the set of entities, each entity in the subset of the set of entities determined to be consistent. The program also provides the subset of the set of entities to the application.

Abstract: Computer-readable media, methods, and systems are disclosed for automatic generation of dynamic application trace logs associated with a running application. A log viewer presents application log entries associated with an application execution log having been generated in connection with a previous execution of the running application. The application execution log is analyzed to identify application execution log context descriptors. The application execution log context descriptors are extracted from the application execution log. The application execution log context descriptors are transmitted to the running application. Matching templates that match each of the one or more application execution log context descriptors are received from the running application. The tracing templates that have an associated context relevance score are received from the running application.

Abstract: Some embodiments provide a non-transitory machine-readable medium that stores a program executable by at least one processing unit of a device. The program receives a set of entities for updating a set of master data stored on a set of storages. The program further determines, for each entity in a subset of the set of entities, whether the entity exists in the set of master data stored on the set of storages. Based on the determinations, the program also updates the set of master data stored on the set of storages with the set of entities.

Abstract: Techniques and solutions are provided for executing database triggers. In particular, disclosed techniques allow for the creation of database triggers with multiple insert statements. For a trigger that includes first and second insert statements, first and second tables are created, respectively for the first and second insert statements, that each include a sequence column. At least the first insert statement references a sequence that is incremented during trigger execution. The sequence columns of the first and second tables have a number of elements corresponding to a number of times a given insert operation will execute as a result of a database operation that satisfies the conditions of the trigger. The first and second insert operations are executed using the respective first and second tables.

Abstract: The present disclosure relates to computer-implemented methods, software, and systems for managing asynchronous data replication in a multiple availability zone cloud environment. Metadata for files for asynchronous replication at a second availability zone is stored at an in-memory data grid of a first instance of a storage service at a first availability zone at a multiple availability cloud platform that provides storage services. The in-memory data grid includes a queue data structure of metadata records and a map of metadata records. In response to determining that connection from the first availability zone to the second availability zone is available, asynchronous data replication for files identified at the map is executed. A file for replication is identified at the map and provided for replication at a second file storage at the second availability zone through a replication interface of a second instance of the storage service at the second availability zone.

Abstract: Intelligent annotation of data models can be implemented. In one embodiment, the method can receive a data model including entities and relationships between the entities. An entity can include a set of attributes. The method can annotate the data model by defining a logical entity including one or more of the entities, validate the logical entity, and automatically generate a message structure corresponding to the logical entity. The message structure can include properties mapped to at least some of the attributes of entities contained in the logical entity. The properties can be defined in one or more database tables represented by the logical entity.

Abstract: Techniques and solutions are provided to facilitate monitoring and maintaining a computing environment. A current state of one or more parameters, resources, or services of a computing environment is determined. The current state is compared with predefined state information. It is determined that the current state is inconsistent with the predefined state information. It is determined whether an update condition is present. If the update condition is not present, the computing environment is automatically updated such that the current state matches the predefined state information. If the update condition is present, the predefined state information is updated.

Abstract: A first version of a chatbot using natural language processing conducts conversations with a plurality of users. The chatbot provides responses by triggering a plurality of skills including a fallback skill that is triggered when no other skill corresponds to the intent of the user. A second version of the chatbot is deployed while the first version of the chatbot is concurrently engaged in a set of conversation sessions. The second version of the chatbot is configured to trigger other skills besides the fallback skill for at least a portion of the conversations in which the first version of the chatbot triggered the fallback skill. New conversation sessions are mapped to the second version of the chatbot while the set of conversation sessions the first version of the chatbot is engaged in are still mapped to the first version of the chatbot.

Abstract: Techniques for validation against aggregation across different versions of data obtain a selection of a data set, a dimension configuration for a data visualization, and a filter configuration for the data visualization. Then it is determined whether the data visualization is valid based on a determination that the zero or more filter dimensions include the version dimension, or a determination that the zero or more selected dimensions include the version dimension, or a determination that the zero or more selected dimensions include at least one of the measurable dimensions and that all measurable dimensions of the selected dimensions are restricted to a single version of data from among the multiple versions of the same data. The data visualization is generated in response to a determination that the data visualization is valid.

Abstract: Systems and processes for aligning weakly-annotated data to recognized characters in a document are provided. In a method for aligning annotation data to recognized characters, annotation words and character recognition tokens are received, and a search algorithm is performed to align the annotation words to the tokens in a stepwise manner. At each step, an annotation word is aligned to one or more tokens, and a cost of each respective alignment is calculated. Once all annotation words are aligned, a full set of annotation word-token pairs corresponding to the annotation is selected based on a total cost of alignment for that set. A bounding box enclosing the tokens in the selected full set is generated and output to a target application.

Abstract: Methods, systems, and computer-readable storage media for a multimodal machine learning image and text combined search method. One example method includes processing items that each have an associated image and a textual description. A first image feature vector is generated by processing a first image using a first machine learning model. A first textual feature vector is generated by processing a first textual description using a second machine learning model. The first image feature vector and the first textual feature vector are combined to generate a first combined feature vector for a first item. Similarity lists of similar items are generated for the first item based on similarities between the first image feature vector, the first text feature vector, the first combined feature vector and respective corresponding vectors of other items. The similarity lists for the first item are combined to generate a combined similarity list for the first item.

Abstract: A system and method including receiving, from a first user of a first service, an indication of a second service to integrate with the first service; correlating a presence of the first user of the first service with an identifier of the first user in the second service; receiving, from a second user of the first service, an indication of the second service to integrate with the first service; correlating a presence of the second user of the first service with an identifier of the second user in the second service; receiving, from the second service via an application programming interface, a replication of a statement of work generated by the second service and associated with the second user; and persisting the replication of the statement of work in a data store of the first service that is accessible by the first user of the first service.

Abstract: Provided is a system and method for training and validating models in a machine learning pipeline for failure mode analytics. The machine learning pipeline may include an unsupervised training phase, a validation phase and a supervised training and scoring phase. In one example, the method may include receiving an identification of a machine learning model, executing a machine learning pipeline comprising a plurality of services which train the machine learning model via at least one of an unsupervised learning process and a supervised learning process, the machine learning pipeline being controlled by an orchestration module that triggers ordered execution of the services, and storing the trained machine learning model output from the machine learning pipeline in a database associated with the machine learning pipeline.